Canva 发布首个自研图像模型 Craft
原标题：Canva launches Craft, its first in-house image model
AI 摘要
Canva 发布其首个自研图像生成模型 Canva Craft，由内部实验室 Canva Research 从零训练，主要基于数百万设计模板而非照片。该模型先生成构图、比例和文字，再输出最高 2K 分辨率、30 多种宽高比的图像，支持 55 种语言文字渲染，并可通过 Magic Layers 拆分为可编辑图层。Canva 称其在 DesignArena 排名第四、人像与风景类第一，已通过 Dream Lab 和公共 API 向所有用户开放，并计划后续接入 ChatGPT、Gemini 等聊天机器人（MCP 集成）。
正文节选
Canva launches Craft, its first in-house image model Design platform Canva has released Canva Craft, its first in-house image generation model. The company says it can compete with comparable models from Google, OpenAI, ByteDance, and Meta. According to Canva, its internal lab Canva Research trained the model from scratch, relying mainly on millions of design templates rather than photos. Craft first plans out image composition, proportions, and text, then generates images at up to 2K resolution